Non steroidal anti-inflammatory drugs are among the most promising classes of drugs for the prevention and possibly treatment of Alzheimer's disease (AD). A rapidly increasing number of large-scale therapeutic trials of such drugs are being initiated. The most likely target of NSAIDs in the brain is cyclooxygenase (COX)-2. We found that COX-2, but not COX-1 expression, is elevated in the neurons of AD brain, where it correlates with amyloid plaque density and neuronal atrophy. In this revised application, using a combination of in situ hybridization and immunocytochemical techniques, we will further study the regional distribution and cell-type expression of COX-2 and other inflammatory markers in the AD brain. To determine the relationship between COX-2 expression in the AD brain and clinical measures of disease activity, COX-2 expression will be correlated with antemortem assessment of dementia. Because therapeutic trials of potential disease-modifying regimens select patients at one or more stages of clinical disease, these studies will determine the relationship between AD clinical stage and COX-2 expression. In parallel studies the effect of COX inhibitors on COX-2 mediated responses in the brain will be explored using a transgenic mouse model of human (h)COX-2 overexpression in neurons. In preliminary studies using primary neuron cultures derived from these transgenic mice, we found that hCOX-2 overexpression potentiates beta amyloid (Abeta) neurotoxicity in vitro through potentiation of oxidative stress mechanisms. We will use this model system to compare the neuroprotective activity of various COX inhibitors on Abeta toxicity in vitro, and to study the mechanism of such neuroprotection. Based in part on the outcome of these studies, we will then test the brain activity of NSAID regimens administered systemically. We have established in our preliminary studies that transgenic mice with neuronal overexpression of hCOX-2 show increased lipid peroxidation in brain as measured by levels of malondialdehyde (MDA) along with elevated prostaglandin (PG)F2alpha. Preliminary studies also indicate increased expression of components of the complement cascade in the brain of hCOX-2 transgenics. Based on our evidence that COX-2 in neurons is indeed the appropriate target for NSAID regimens in AD, this transgenic model provides a unique method of measuring relevant brain activity of COX inhibitors. The outcome of the proposed studies will be immediately relevant to the design of human trials.